A variety of products may be fabricated by thin film processes. Examples of the products that may be fabricated by the deposition of thin film materials include interferometer stacks for optical control and solar control. An example of a solar control product is disclosed in U.S. Pat. No. 5,494,743 to Woodard, et al entitled "ANTIREFLECTION COATINGS", the disclosure of which is incorporated herein by reference. More specifically, Woodard, et al disclose a polymeric substrate having anti-reflective coatings disposed thereon. The anti-reflective coatings consist of one or more inorganic metal compounds with indices of refraction higher than that of the polymeric substrate.
Thin film materials that are used for optical control are generally comprised of a series of layers of metals and dielectrics of varying dielectric constants and indices of refraction. These thin film materials may be used, for example, to reduce glare or reflection. Thin film materials may also be used as solar control films for low emission of infrared radiation in order to reduce the loss of heat.
In the manufacture of thin film materials for optical control, many interferometer stacks will have a top layer of silicon dioxide. An antireflective layer for a single layer of material having an index of refraction greater than 1.00 will have an index of refraction equal to the square root of the index of refraction of the single layer material. The thickness of the material calculated at the center wavelength of the frequency band at issue, more precisely, the optical thickness is 1/4 of the wavelength at the center frequency. For example, the human eye generally sees light having a wavelength between 4000 .ANG. and 7000 .ANG.. Therefore, the thickness of the optical coating for anti-reflection at 5500 .ANG. is about 1375 .ANG.. Optical properties including the index of refraction and transparency as well as with the mechanical properties of silicon dioxide make it the material of choice for anti-reflective coatings.
A number of processes are currently utilized to deposit thin film materials, some of which are described in Thin Film Processes, John L. Vossen and Warner Kern, eds., Academic Press, Inc., New York, N.Y., 1978. The fundamentals of chemical vapor deposition are disclosed in Chapter III-2 of Thin Film Processes by Warner Kern and Vladimir S. Ban. Chemical vapor deposition, CVD, as a method of forming and depositing material causes the constituents of a gas or vapor phase of a material to form a product which is deposited on some surface. Therefore, the chemical reaction may be either endothermic or exothermic.
The reactants of a CVD process are the logical result of the stack design and are determined by the precursor materials. For example, if silicon dioxide (SiO.sub.2) is desired to be deposited, silane (SiH.sub.4) may be oxidized by oxygen (O.sub.2) to yield silicon dioxide as the desired product and a by-product of hydrogen (H.sub.2). Alternatively, silane may be decomposed to deposit an amorphous silicon alloy material on a substrate. For example, products may be formed by energizing the reactants to a reaction temperature. The reaction temperature may be achieved by any suitable method known in the art including R.F. glow discharge and electrical resistive heating. A CVD reaction may occur in a wide range of pressures from above an atmosphere to a less than a millitorr.
Low pressure CVD processes offer substantial advantages over CVD processes operating at about atmospheric pressure. The diffusity of a gas and the mean free path of gas molecules is inversely related to pressure. As the pressure is lowered from about atmospheric pressure to 1 torr, the effect is an increase of approximately 2 orders of magnitude in the diffusion constant. Commonly assigned, U.S. Pat. Nos. 4,517,223 and 4,504,518 to Ovshinsky, et al both entitled "METHOD OF MAKING AMORPHOUS SEMICONDUCTOR ALLOYS AND DEVICES USING MICROWAVE ENERGY", the disclosures of which are incorporated herein by reference, described processes for the deposition of thin films onto a small area substrate in a low pressure, microwave glow discharge plasma. As specifically noted in these patents, operation in low pressure regimes not only eliminates powder and polymeric formations in the plasma, but also provide the most economic mode of plasma deposition.
A low pressure microwave initiated plasma process for depositing a photoconductive semiconductor thin film on a large area cylindrical substrate using a pair of radiative waveguide applicators in a high power process is disclosed in commonly assigned, U.S. Pat. No. 4,729,341 to Fournier, et al for "METHOD AND APPARATUS FOR MAKING ELECTROPHOTOGRAPHIC DEVICES", the disclosure of which is incorporated herein by reference. However, the principles of large area deposition described in the '341 patent are limited to cylindrically shaped substrates and the teachings provided therein are not directly transferable to an elongated web of substrate material.
The use of a microwave radiating applicator has been extended to chemical vapor deposition onto an elongated web of substrate material in commonly assigned U.S. Pat. No. 4,893,584 to Doehler, et al for "LARGE AREA MICROWAVE PLASMA APPARATUS", the disclosure of which is incorporated herein by reference. By optimizing the isolating window to withstand compressive forces, the thickness of the window may be minimized to provide for rapid thermal cooling, whereby the '584 patent achieves a high power density without cracking the window. Furthermore, by maintaining the apparatus of the '584 patent at subatmospheric pressures, it is possible to operate the apparatus at a pressure approximating that required for operation near the minimum of a modified Paschen curve. As disclosed in commonly assigned U.S. Pat. No. 4,504,518, a Paschen curve is the voltage needed to sustain a plasma at each pressure. A modified Paschen curve is related to the power required to sustain a plasma at each pressure. The normal operating range is dictated by the minimum of the curve. Additionally, the low pressures allow for a longer mean free path of travel for the plasma species, thereby contributing to overall plasma uniformity.
In a CVD process, a sufficient proportion of feedstock gases are provided to achieve a correct stoichiometric deposition of materials. An excellent method for chemical vapor deposition is disclosed in commonly assigned U.S. Pat. No. 5,411,591 to Izu, Dotter, Ovshinsky, and Hasegawa entitled "APPARATUS FOR THE SIMULTANEOUS MICROWAVE DEPOSITION OF THIN FILMS IN MULTIPLE DISCRETE ZONES", the disclosure of which is incorporated by reference herein, Izu, et al disclose an apparatus for the microwave plasma enhanced chemical vapor deposition of thin film material onto a web of substrate material utilizing a linear microwave applicator. By maintaining the plasma region at subatmospheric pressures, a longer mean free path of travel for the plasma species is available, which contributes to the overall plasma uniformity.
In order to maintain a uniform plasma over a much wider substrate, about 1 meter or wider, spacing between the windows must be decreased. As the spacing between the windows of the linear applicator decrease, the potential for shorting increases. It is not possible to maintain a plasma if the linear applicator is prone to shorting. One advantage of a CVD process is the film deposition rate. The product formation rate in a CVD apparatus is related to the flow rate of the feedstock gases. As the rate of product formation increases, the deposition rate also increases. So long as enough energy is provided to react the feedstock gases, the deposition rate is limited by the rate at which non-deposited species may be evacuated from the CVD apparatus. Although a CVD process works well for many thin film materials, there are many materials which are desired and cannot be deposited by any known CVD process, such as indium tin oxide, ITO.
Another known method of depositing thin film material is a PVD (Physical Vapor Deposition) process. There are a number of PVD processes known in the art of thin film material deposition, many of which are disclosed in THIN FILM PROCESSES. J. L. Vossen and W. Kern, eds., Ch. II, Academic Press, New York, N.Y. 1978.
A common PVD process is sputtering which deposits fine particles from a source material. Although the nomenclature is unintuitive, the source of the material to be deposited upon the substrate is called the target. The term "target" evolves from the process of bombarding the source material with a charged noble gas. The target is affixed to a cathode which is a plate having a negative electrical bias. The target faces the substrate material which may be grounded, floating, biased, heated, cooled or some combination thereof. An inert reaction gas, typically argon, is introduced and ionized to provide a medium for transporting an electrical charge. The reaction gas may be ionized by a number of methods including an anode plate, a positively biased inlet port or by biasing the substrate itself. The positively charged reaction gas ion is repelled from the positively charged source and is electrically attracted to the target plate where the positively charged ion strikes the target and removes target atoms by momentum transfer. The removed atoms travel toward the substrate where they condense into thin films.
Although a sputtering process generally does not consume gas for purposes of thin film deposition, it is desirable to flow an inert gas. Flowing the inert gas provides for the removal of impurities which may otherwise accumulate within the chamber. When flowing the inert gas, a pumping scheme should be employed in order to maintain the pressure within the sputtering chamber. It is important to maintain an isobaric condition in the vicinity of the sputtering targets. A pressure gradient will result in a nonuniform bombardment of the sputtering target and consequently non-uniform film deposition. Generally, the chamber pressure for sputtering processes is 75 millitorr or lower. Low pressure sputtering, where the sputtering chamber pressure is about 10 millitorr or less, provides reaction gas ionization far away from the cathode where the chance of the electrical charge being lost to the chamber walls is greatly increased. Therefore, ionization efficiencies are low and self-sustained discharges cannot be maintained in a planar sputtering process.
Reactive sputtering, a method that may be used to form oxides for example, is conducted at a very low pressure, about 5 mtorr or less. The goal in reactive sputtering is to increase the amount of gas phase chemistry, which will increase the probability of collisions, which may be achieved by raising the pressure. However, if SiO.sub.2 is to be deposited by DC sputtering for example, a silicon target is used in an atmosphere containing oxygen. However, oxygen will react with the silicon target material, forming SiO.sub.2, which is an insulator. A DC current cannot be maintained in the present example once the silicon target is oxidized; the charged particle will not have an electrical field to move through.
By the addition of a magnetic field to a sputtering process, sputtering can be maintained at a pressure below 10 millitorr. The mean free path of a charged particle is increased by the addition of the magnetic field. By applying a magnetic field perpendicular to an electric field, the path of the electron is influenced and becomes perpendicular to both the magnetic field and the electrical field. A planar magnetron sputtering device, for example, having a plurality of permanent magnets which are disposed parallel to one another and oriented with alternating polarity on one plane, creates a circular or oval electron path. With the addition of an electrical field, the charged particle takes on a helical path.
The helical path of a charged particle has two advantages: first, the charged particle is prevented from contacting the chamber walls by the presence of the magnetic field, thereby increasing low pressure efficiency; and second, by increasing the length of the traveled path, the potential for collision with other particles has increased.
Although sputtering is a common and well-refined practice, it does have some disadvantages. One of the disadvantages associated with sputtering is the rate of deposition. For example, silicon dioxide can be deposited by both magnetron sputtering and microwave plasma enhanced chemical vapor deposition. The deposition rate of silicon dioxide for pulsed magnetron sputtering is 10-20 .ANG. per second while silicon dioxide deposited by microwave plasma enhanced chemical vapor deposition is deposited at a rate of 100-200 .ANG. per second, an order of magnitude improvement. However, as noted above, there are materials, such as ITO, for which there are no known methods for deposition by chemical vapor deposition.
Thin film materials for the manufacturer of interferometer stacks for optical and thermal control generally consist of multiple layers of materials having a determined thickness layered upon a substrate. The materials and their associated thickness' are collectively referred to as a "stack." A stack is designed to achieve a particular purpose, whether that purpose be optical control, solar control or any other design objectives sought to be achieved. As mentioned above, many optical and solar control stacks have a relatively thick, about 1000 .ANG., top layer of SiO.sub.x. If a stack requires at least one layer to be sputtered, then one of two alternatives is available, under the current state of the art, to produce the top layer of SiO.sub.x. The first alternative is to sputter the entire stack. However, because of the sputtering deposition rate of SiO.sub.x and the required material thickness of SiO.sub.x for the top layer of the stack, a substantially long process time is required to manufacture the stack. Alternatively, all layers except for the top layer of SiO.sub.x may be manufactured by PVD and then the entire roll of sputtered substrate material is transported to a machine for CVD of the 1000 .ANG. layer of SiO.sub.x. Although both of these approaches create the desired final product, the time required to manufacture the stack is substantially long, resulting in higher production costs and reduced efficiency. Furthermore, if the coating is intended for a wide material, about 1 meter, such as a window for a commercial building, the state of the art does not provide a means for depositing a uniform layer of material by CVD.
Therefore, there exists a need in the art for an apparatus which substantially reduces the amount of time required to manufacture a product consisting of multiple layers of thin film material deposited on a substrate by including a PVD process and CVD process in single machine.
Furthermore, there exists a need in the art for a CVD process that is capable of depositing a uniform layer of material onto a widened substrate.